Active vs passive debate again... 
Passives XOs have practical advantages like simplicity, low part count, you only need X amplifier channels for X amount of speakers.
But... Amplifier damping can be severely compromised in favour of a shaped frequency response. In other words, once the response is a few dB down, the damping factor is likely to be completely shot, and resonant box modes will reverberate back through the speaker cone. Notch filters might improve things by short-circuiting the speaker in a given frequency area, but then you need to protect the amplifier by adding additional passives in series.
People forget that dynamic speakers are transducers. Speakers not only make sound from electric current, but they also generate spurious currents whenever something pushes the voice coil, such as panel vibrations, resonant air modes inside the box, or cone break-up modes. Maybe a well-designed passive XO takes all that into account, but they seem full of compromises. Even so-called parallel designs usually share the same amplifier channel, which has a non-zero output impedance. So if the midwoofer in a 2-way design generates some spurious current, then unless it's hit dead-on by a notch filter and 100% short-circuited, a voltage will be generated, which will affect the tweeter even though it's a parallel XO.
Active designs avoid all that nonsense by separating the channels and keeping each speaker completely short-circuited to the amplifier at all frequencies. I'm not saying that this is ideal -- perhaps the electrical damping could be fine-tuned in an attempt to critically damp some cone resonances, but Qtc usually exceeds 0.5 in the bass, so any gains with softer break-up modes are likely to be offset by worse box modes.
Passives XOs have practical advantages like simplicity, low part count, you only need X amplifier channels for X amount of speakers.
But... Amplifier damping can be severely compromised in favour of a shaped frequency response. In other words, once the response is a few dB down, the damping factor is likely to be completely shot, and resonant box modes will reverberate back through the speaker cone. Notch filters might improve things by short-circuiting the speaker in a given frequency area, but then you need to protect the amplifier by adding additional passives in series.
People forget that dynamic speakers are transducers. Speakers not only make sound from electric current, but they also generate spurious currents whenever something pushes the voice coil, such as panel vibrations, resonant air modes inside the box, or cone break-up modes. Maybe a well-designed passive XO takes all that into account, but they seem full of compromises. Even so-called parallel designs usually share the same amplifier channel, which has a non-zero output impedance. So if the midwoofer in a 2-way design generates some spurious current, then unless it's hit dead-on by a notch filter and 100% short-circuited, a voltage will be generated, which will affect the tweeter even though it's a parallel XO.
Active designs avoid all that nonsense by separating the channels and keeping each speaker completely short-circuited to the amplifier at all frequencies. I'm not saying that this is ideal -- perhaps the electrical damping could be fine-tuned in an attempt to critically damp some cone resonances, but Qtc usually exceeds 0.5 in the bass, so any gains with softer break-up modes are likely to be offset by worse box modes.
Regarding the damping factor…
Quote
To put it simple: In passive loudspeakers, the frequency-separating filters are located in
between the amplification and the driver units and necessarily cause a loss of precision with which the amp can control the drivers.
The best (and therefore quite expensive) crossover networks available today can reduce this
impairment but never avoid it. Even the best ones diminish the damping factor drastically,
on average down to five (yes, 5!) percent of the original value!
Unquote...The extract from Definition Active vs.Passive pdf :Home Audio Downloads | ADAM Audio GmbH
Quote
To put it simple: In passive loudspeakers, the frequency-separating filters are located in
between the amplification and the driver units and necessarily cause a loss of precision with which the amp can control the drivers.
The best (and therefore quite expensive) crossover networks available today can reduce this
impairment but never avoid it. Even the best ones diminish the damping factor drastically,
on average down to five (yes, 5!) percent of the original value!
Unquote...The extract from Definition Active vs.Passive pdf :Home Audio Downloads | ADAM Audio GmbH
The only place where a high damping factor is useful is close to driver fs, otherwise you'd have more use of having an extremely low damping factor, say 0.001. Then you don't have to worry about impedance variation distortion in the drivers anymore which is nice.
Assuming low output impedance / high damping factor on the amp just makes it easier to design the crossover, but not necessarily better.
Assuming low output impedance / high damping factor on the amp just makes it easier to design the crossover, but not necessarily better.
It's worth noting, that now 4 pages into this thread, not one person has supported their claims with any refutable evidence as outlined below. Some of the opinions are fine, we're all entitled to such, but if people are going to make seemingly technical claims in stating that their beliefs are scientifically proven, they should take the initiative to support it as such. What I have seen in this thread, like many on this site, are empty statements. They might impress some, but as a critical thinker I am poised to question their validity.
Does active offer better dynamic signal fidelity, or just steady-state sinusoidal tones? If a passive crossover adds more distortion, what type is it and how severe is is when compared to an active filter? How comprehensive were the results that people base their beliefs on? Why do they never post their measurements? Does active function better for high or low pass filtering? Could the two approaches work together to better effect? How much noise does active filtering add? What happens to the high-frequency drivers if the active filter fails in operation or the amplifier presents a momentary direct current while there is no high-pass passive filter inplace? What if the user, in all his good intentions, creates an unstable biquad when using active filters? I can think of many more questions to add to these, but they are rhetorical. I guess that I'm taken aback that so many have such determined beliefs but have not adressed the salient issues at hand. I know I'm certainly not interested in links or articles. I want to see hard evidence from people in this thread who claim to know so much.
Otherwise, ten pages from now people will still be making empty claims and trying to confirm their personal biases with excerpts from so-called experts and online links, and this thread will fall among the redundant. The post below highlights what this thread, and its claims, requires to gain a solid footing for further considerations.
Does active offer better dynamic signal fidelity, or just steady-state sinusoidal tones? If a passive crossover adds more distortion, what type is it and how severe is is when compared to an active filter? How comprehensive were the results that people base their beliefs on? Why do they never post their measurements? Does active function better for high or low pass filtering? Could the two approaches work together to better effect? How much noise does active filtering add? What happens to the high-frequency drivers if the active filter fails in operation or the amplifier presents a momentary direct current while there is no high-pass passive filter inplace? What if the user, in all his good intentions, creates an unstable biquad when using active filters? I can think of many more questions to add to these, but they are rhetorical. I guess that I'm taken aback that so many have such determined beliefs but have not adressed the salient issues at hand. I know I'm certainly not interested in links or articles. I want to see hard evidence from people in this thread who claim to know so much.
Otherwise, ten pages from now people will still be making empty claims and trying to confirm their personal biases with excerpts from so-called experts and online links, and this thread will fall among the redundant. The post below highlights what this thread, and its claims, requires to gain a solid footing for further considerations.
Last edited:
I have done some approaches for transient-improved active crossovers. One with a Manger driver and a 8" Woofer and the other one was a 8" coaxial. I could post step responses of those. But I don't have the step repsonse of a passive version to compare because:
1.) I wasn't interested at all in trying a passive one and
2.) The same couldn't have been achieved passively.
For both of them some computing capability was necessary, it doesn't matter wheter digital or analog.
Regards
Charles
1.) I wasn't interested at all in trying a passive one and
2.) The same couldn't have been achieved passively.
For both of them some computing capability was necessary, it doesn't matter wheter digital or analog.
Regards
Charles
Last edited:
Having moved from an analog active system to a digital active system a few months ago, I will offer this comment:
Analog filters depend on the precision of the parts to get what is calculated. It can be difficult to purchase capacitors with better tolerance than +- 20% for many values. Digital crossovers, however, don't depend on component values. Can you hear the difference? I was certainly able to measure the difference with a scope and DMM. Over 4 db around the crossover range.
Analog filters depend on the precision of the parts to get what is calculated. It can be difficult to purchase capacitors with better tolerance than +- 20% for many values. Digital crossovers, however, don't depend on component values. Can you hear the difference? I was certainly able to measure the difference with a scope and DMM. Over 4 db around the crossover range.
Try to hang on to your sense of humor Scott, in my book that's a mild leg-pull. When I insult somebody, they know it. Hardly surprising that your reply evades the main thrust of my post.
Most issues in engineering are absolutely cut-and-dried, to the point where we can state the margin of error, it's only at the bleeding edge that things get vague. This certainly isn't the bleeding edge.
There was absolutely no reason to disagree with me in the first place, I haven't said anything remotely controversial.
The range of freedoms available to the designer of practical active systems is a superset of the range of freedoms available to the designer of practical passive systems, due to the physical constraints already agreed. And they waste less power. And they require less voltage.
Now, it seems that I have stepped into a controversy. I wasn't aware that there was an active/passive debate, and I don't see myself as a 'supporter' of either type of system. I have no active systems at present, but if I were to build a system, it would be an active one, since I regard passive types as legacy systems and I have a very good book on the subject which eases the choice and design of opamp filters, but I would not rule out a hybrid, nor the use of DSP for the xover, it's a question of cost and the practicality of construction, since, in my experience, electronics that is pretty much beyond reproach can be expected these days. Except of course in so-called 'high-end' audio systems.
Everybody wants to be an artist. In this I'm lucky, I'm a musician as well as an engineer, nobody disputes that musicians can be artists. When engineers start to tell me they're artists, they stretch my credulousness to the point of nausea. Anybody would imagine that compromise is a desirable quality in an engineers character. You must have spent too much time on social media sites.
kouiky
Almost everything in electronics is calculated from first principles. We can reach meaningful conclusions without the benefit of measurement, purely by the application of basic physics, maths and logic. If you do not understand this then your contribution to the discussion is only noise.
Most issues in engineering are absolutely cut-and-dried, to the point where we can state the margin of error, it's only at the bleeding edge that things get vague. This certainly isn't the bleeding edge.
There was absolutely no reason to disagree with me in the first place, I haven't said anything remotely controversial.
The range of freedoms available to the designer of practical active systems is a superset of the range of freedoms available to the designer of practical passive systems, due to the physical constraints already agreed. And they waste less power. And they require less voltage.
Now, it seems that I have stepped into a controversy. I wasn't aware that there was an active/passive debate, and I don't see myself as a 'supporter' of either type of system. I have no active systems at present, but if I were to build a system, it would be an active one, since I regard passive types as legacy systems and I have a very good book on the subject which eases the choice and design of opamp filters, but I would not rule out a hybrid, nor the use of DSP for the xover, it's a question of cost and the practicality of construction, since, in my experience, electronics that is pretty much beyond reproach can be expected these days. Except of course in so-called 'high-end' audio systems.
Everybody wants to be an artist. In this I'm lucky, I'm a musician as well as an engineer, nobody disputes that musicians can be artists. When engineers start to tell me they're artists, they stretch my credulousness to the point of nausea. Anybody would imagine that compromise is a desirable quality in an engineers character. You must have spent too much time on social media sites.
kouiky
Almost everything in electronics is calculated from first principles. We can reach meaningful conclusions without the benefit of measurement, purely by the application of basic physics, maths and logic. If you do not understand this then your contribution to the discussion is only noise.
Your post is false, because without measurements you bring only opinions and views, but no facts. This in itself makes your post falliable and no better than any others. At this time, you have yet to support any of your claims with proof, and have contributed nothing of value beyond heresay and supposition under the guise of scientific fact. If it cannot be supported, repeated or proven, then it is not real science.
Last edited:
And this explanation is supposed to be what, serious science?
Manufacturers have to make up reasons to make people believe
time has come to buy something else. Passives are history, long
live the actives!
Here is the measured step response of my 3-way (15" woof, horn mid and tweet) with passive XO at the LP:
Same speaker, but measured step response using 3 way digital XO (same XO points as the passive above) and time alignment at the LP:
For comparison, this is the electrical step response of my Lynx Hilo AD DA converter in external loopback mode:
All measures produced using REW's default settings with a 10Hz to 24 kHz sweep and 500ms window.
Some background info on why use step response to measure a speakers time coherence:
Measuring Loudspeakers, Part Two Page 2 | Stereophile.com
Measuring Loudspeakers, Part Two Page 3 | Stereophile.com
Long and boring article, with measurements, on how I achieved this:
Computer Audiophile - Advanced Acourate Digital XO Time Alignment Driver Linearization Walkthrough
The only differences from the article are that I changed out the HF horn and driver for a QSC CD waveguide mated to a BMS 4540 driver and the above measures do not include driver linearization.
Hope that helps. Cheers, Mitch
Same speaker, but measured step response using 3 way digital XO (same XO points as the passive above) and time alignment at the LP:
For comparison, this is the electrical step response of my Lynx Hilo AD DA converter in external loopback mode:
All measures produced using REW's default settings with a 10Hz to 24 kHz sweep and 500ms window.
Some background info on why use step response to measure a speakers time coherence:
Measuring Loudspeakers, Part Two Page 2 | Stereophile.com
Measuring Loudspeakers, Part Two Page 3 | Stereophile.com
Long and boring article, with measurements, on how I achieved this:
Computer Audiophile - Advanced Acourate Digital XO Time Alignment Driver Linearization Walkthrough
The only differences from the article are that I changed out the HF horn and driver for a QSC CD waveguide mated to a BMS 4540 driver and the above measures do not include driver linearization.
Hope that helps. Cheers, Mitch
Not surprising snce we do not yet know how to do "a full set of complete measurements".
dave
This is not fact. I know of no way to do a series XO actively. So that is part of the passive XO set that is not part of the active XO set. I am sure there are other examples.
Not specifically electronics, but only true for the ~5% of problems that do not become chaotic.
dave
Having a negligible driving impedance is not necessarily superior, for the following reasons:
Current driven VCs do not suffer thermal compression.
Positional coil inductance modulation does not create distortion with current drive.
Magnetic eddy current generation does not create distortion with current drive.
Also, an advantage in some cases would be a slight extension in frequency response at potentially both ends of the loudspeaker spectrum.
Also, the active range of xover options is not a 'superset' of the passive range, as I have already proved.
Current driven VCs do not suffer thermal compression.
Positional coil inductance modulation does not create distortion with current drive.
Magnetic eddy current generation does not create distortion with current drive.
Also, an advantage in some cases would be a slight extension in frequency response at potentially both ends of the loudspeaker spectrum.
Also, the active range of xover options is not a 'superset' of the passive range, as I have already proved.
Last edited:
Nice example of the pre-ringing you get with a typical digital XO (circled in yellow).
Note: one may get the idea that i am a proponent of passive XOs. Not true, i am more of an agnostic, preferring no XO. If i have to do an XO i tend towards line level, but as much because doing a decent passive XO is hard.
dave
Attachments
A little spice or spot of pink salt can add just the right amount of sparkle to a nice sauce, but could there be more than a dash of tautology in this "conversation"?
Not being an engineer, I can only comment on my personal experience with over 40yrs of audio -"there is no single right way to achieve any goal" - now that's either a scientific fact, or just one idiot's opinion - you choose.
If counting the system in the workshop, I currently have 5 systems that I listen to regularly, 2 of which are conventional passive XO, and the others are single driver Fullrangers - including the 5.1 Surround system. As these nowadays all include substantial flexibility for bass management - my Onkyo will even accommodate digital XO and bi-amping of front mains- I think it's reasonable to categorize that as a fully active system.
For me each one offers acceptable performance compromises.
Not being an engineer, I can only comment on my personal experience with over 40yrs of audio -"there is no single right way to achieve any goal" - now that's either a scientific fact, or just one idiot's opinion - you choose.
If counting the system in the workshop, I currently have 5 systems that I listen to regularly, 2 of which are conventional passive XO, and the others are single driver Fullrangers - including the 5.1 Surround system. As these nowadays all include substantial flexibility for bass management - my Onkyo will even accommodate digital XO and bi-amping of front mains- I think it's reasonable to categorize that as a fully active system.
For me each one offers acceptable performance compromises.
Nice example of the pre-ringing you get with a typical digital XO (circled in yellow).
Note: one may get the idea that i am a proponent of passive XOs. Not true, i am more of an agnostic, preferring no XO. If i have to do an XO i tend towards line level, but as much because doing a decent passive XO is hard.
dave
Um, that is the step response of my Lynx Hilo AD DA converter only. There is no digital XO or anything else in the signal path for that measure... The signal path is REW digital output ->Hilo DA conversion->wired loopback->Hilo AD conversion->REW digital input->REW display.
It is still pre-ringing. something that is very unnatural, and hence will stand-out to the ear-brain system at much lower levels than something that can be found in nature.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.